基于周期分解的ARIMA模型在甲肝发病率预测中的应用

摘要:目的　探讨基于周期分解的ARIMA模型在我国甲肝月发病率预测中的应用，并比较其与SARIMA模型的预测效果。方法　收集2004年1月-2014年12月我国甲肝月发病率资料，用SPSS13.0分别拟合两种模型，并用2014年的数据评价模型的预测效果。结果　基于周期分解的ARIMA模型的拟合及预测的MRD，MER，MSE和MAE分别为4.4691，0.0446，0.0002，0.0092；4.1310，0.0415，0.0001，0.0066。SARIMA模型的拟合及预测的MRD，MER，MSE和MAE分别为7.2979，0.0781，0.0003，0.0185；6.4407，0.0708，0.0002，0.0110。结论　基于周期分解的ARIMA模型拟合和预测效果优于SARIMA模型。它可以提高预测的精度，具有较好的应用价值。

Application of ARIMA model based on seasonal decomposition in prediction on incidence of Hepatitis A,China

Abstract:Objective To explore the application of ARIMA model based on seasonal decomposition and SARIMA model in the prediction on incidence of hepatitis A in China and compare the predict effect among them. Methods The data of monthly incidence of Hepatitis A from January 2004 to December 2014 in China was collected and SPSS13.0 was used to fit two models respectively. At the same time, The monthly data in 2014 was used to evaluate the effect of prediction. Results The MRD, MER, MSE and MAE fitted and predicated by ARIMA model based on seasonal decomposition were 4.4691, 0.0446, 0.0002, 0.0092 and 4.1310, 0.0415, 0.0001, 0.0066, respectively. The MRD, MER, MSE and MAE fitted and predicted by SARIMA model were 7.2979, 0.0781, 0.0003, 0.0185 and 6.4407, 0.0708, 0.0002, 0.0110, respectively. Conclusion ARIMA model based on seasonal decomposition is superior to the SARIMA model with a good practical value for significantly improving the forecasting precision.